Excel agrees with me too. Here's a 4th order polynomial that best fits the data. The data here is from Regenstrief's dashboard so it's not identical to what they showed, but very very similar.
To be clear, I don't think they *intended* it to be misleading, but it ended up that way. I get they're trying to emphasize the drop from 5/4 (why exactly does the graph start at 5/1?) to now. But there's a difference between making a graph *clearer* vs. *imply something*
Not to belabor this point any further, but to try and turn it into a good learning moment... one sign of a well-fitted trend line is if the data is about 50% *above* and 50% *below* the trend (i.e. it's around the mean). With that in mind look at the original image again...